Hybrid forecasting model to predict air passenger and cargo in Indonesia

Abstract

Forecasting of air passenger and cargo have a major influence on the master plan of the airport infrastructure development and investment by the civil airline. This research aims to obtain the most accurate predictive value of the air passenger and cargo at three international airports Indonesia, namely, Soekarno Hatta, I Gusti Ngurah Rai, and Juanda Airport. Those international airports are the three largest contributors to the number of air passengers and cargo volumes in Indonesia. This research uses a hybrid forecasting method that combines linear and nonlinear models. The combination of two linear and nonlinear models is able to obtain accurate predictions. The first phase is linear modeling with time series regression model (TSR) and Autoregressive Integrated Moving Average with Exogenous Factor (ARIMAX). In the second phase, the error of the linear model is analyzed by using machine learning methods such as Neural Network (NN) and Support Vector Regression (SVR) to capture nonlinear patterns. There are four hybrid models that be applied and compared, i.e. TSR-NN, TSR-SVR, ARIMAX-NN, and ARIMAX-SVR based on the Mean Absolute Percentage Error (MAPE). The results show that hybrid ARIMAX-NN and TSR-NN give more accurate prediction than hybrid TSR-SVR and ARIMAX-SVR.